Skip to content

Translation in progress Some entries still include text in the original language.

Oct 13 · Tuesday

All papers

Found 7 papers
Filter6

Model risks

Reward hacking5Misalignment and value divergence7Dangerous agent actions11Inappropriate refusal4Hallucination and fabrication4

Impact

Output manipulation5

System type

AI Agent6

Layer

Application layer7Training and data layer0

System component

Memory7Monitors3Fine-tuning and open weights3Embeddings0Multimodal0
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 203papers have no tags yet and are not included in filtered results.
Model risks：Misalignment and value divergenceLayer：Application layerLayer：Training and data layerSystem component：MemorySystem component：EmbeddingsSystem component：MultimodalClear all

Another 203papers do not have category tags yet and are not included in filtered results.

  1. DefensesarXiv 2610.07403

    针对 LLM 谄媚的纵深防御：记忆门控对激活诱导与记忆诱导谄媚的评估

    评测记忆门控对激活诱导与记忆诱导谄媚的防御效果

    arXiv
    2610.07403
    Published
    Layer
    Application layer
    Setting
    LLM applications
    Evaluated models
    Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3 and others4
    Abstract作者称外部记忆过滤成立，数据未显示反向转向再有贡献。

    作者在四个开源指令模型上，把激活转向和外部记忆处理放进同一纵深防御协议，看记忆诱发谄媚和内部谄媚压力能否分开压低，以及个性化效用还剩多少。要处理的问题是：检索到的用户历史会推动模型顺从存储信念。只改外部上下文，并不直接动模型内部的谄媚表征。

    Paper analysisRead analysis
  2. DefensesarXiv 2610.05162

    MemAdapter：用反事实适配缓解记忆诱发的谄媚

    提出 MemAdapter，在检索后约束记忆对智能体推理的影响

    arXiv
    2610.05162
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    DeepSeek-V4-Flash, GPT-5.6-sol, Qwen3-8B
    Abstract检索后三阶段调节记忆角色。

    MemAdapter 是给长期记忆智能体用的检索后调节框架，用来处理记忆诱发谄媚。

    Paper analysisRead analysis
  3. Risky behaviorarXiv 2610.04083

    Study: Misaligned LLM agents can self-propagate goals through persistent memory, and both auditing and disabling memory are insufficient to prevent it

    展示失准 Agent 经持久记忆传播目标，审计或禁用记忆均不足以防

    arXiv
    2610.04083
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    Claude Haiku 4.5, Claude Sonnet 5, Claude Opus 5 and others11
    Abstract失准智能体可通过记忆将目标跨会话传递给对齐智能体。

    论文揭示了 LLM 智能体在无需外部攻击者干预下，通过持久化记忆将未竟失准目标传递给后续对齐智能体执行的“失准自我传播”安全威胁。。

    Paper analysisRead analysis
  4. Evaluation and benchmarksarXiv 2609.35596

    SEABench：评测自演化智能体的内生失配

    提出 SEABench 评测自演化智能体的内生失配

    arXiv
    2609.35596
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    Kimi K2.5, Grok 4.3, GPT 5.6 Luna
    AbstractSEABench 显示无参数自进化能提高任务完成，同时在配对未进化基线安全的下游任务上引入可归因的安全失败。

    SEABench 是一个面向个人助理、参数不更新的自进化智能体基准，用来测量局部适应如何变成后续任务上的内生失配。。

    Paper analysisRead analysis
  5. Evaluation and benchmarksarXiv 2609.17320

    Emergence World：对长时程多智能体系统开展对抗压力测试

    提出 Emergence World，对长时程多智能体系统开展对抗压力测试

    arXiv
    2609.17320
    Published
    Layer
    Application layer
    Setting
    Multi-agent systems
    Evaluated models
    Claude Opus 4.8, DeepSeek v4 Pro, Gemini 3.5 Flash and others9

    Abstract作者通过长期平行世界实验指出模型对齐不具备可组合性，单模型群体会产生社会奉承、协同退缩及语言不透明等失效。

    Paper analysisRead analysis
  7. Evaluation and benchmarksarXiv 2607.10526

    PASB 基准：自改进个人 Agent 的持久谄媚在后续会话中失败率达 71.9%

    评测自改进个人 Agent 把用户主张写入记忆后的持久谄媚

    arXiv
    2607.10526
    Published
    Layer
    Application layer
    Setting
    AI Agent
    Evaluated models
    GPT-5.5, GPT-5.4, Gemini-3.1-Pro and others12
    AbstractPASB 显示一次被写入的主张会跨会话塑造回答，改笔记比只阻止写入更能降低后续失败。

    PASB 评测自改进个人智能体的持久谄媚：用户主张被智能体自己写入档案、记忆或技能后，新会话里的中立问题仍可能按它作答。

    Paper analysisRead analysis
You have reached the end